Photo-Control of Adsorption of Dye Metal Complexes Incorporating Chiral Schiff Base Ligands Containing Azo-Groups on TiO₂

To a methanol solution (15 mL) of 4,4'-((1S,2S)-1,2-diaminoethane-1,2-diyl) dibenzoate (0.072 g, 0.25 mmol),¹⁶ methanol solution (15 mL) of potassium hydroxide (0.014 g, 0.25 mmol) and 2-hydroxy-5-(phenyldiazenyl) benzaldehyde (0.1132 g, 0.5 mmol)¹⁷ was added dropwise, and the resultant mixture was stirred at 313 K for 3 h. Then iron(II) sulfate heptahydrate (0.070 g, 0.25 mmol) was added, and the solution was stirred for 3 h. After the reaction and filtration, the filtrate was concentrated under reduced pressure, to obtain a black brown solid. This crude solid was washed with methanol and n-hexane to give rise to the Fe-L complex. Synthesis of other metal complexes were carried out in almost similar way instead of the corresponding metal sources (copper(II) acetate monohydrate or bihydrate).

Fe-L: Yield 0.156 g (80.9 %). Anal. Found: C, 57.14; H, 3.38; N, 9.16%. Calcd. for C₄₂H₃₀N₆O₈K₂Fe: C, 57.14; H, 3.65; N, 9.52%. IR (KBr, cm^-1) 536 (m), 656 (w), 688 (w), 769 (w), 849 (w), 1018 (w), 1114 (m), 1223 (w), 1301 (m), 1378 (m), 1421 (w), 1536 (w), 1606 (s, C=N), 1721 (m, C=O), 2342 (w), 2359 (w), 2951 (w), 3434 (br, s, -OH). μ_eff=4.86 B.M. at 300 K (theoretical value high spin d⁶ is 4.90 B.M.)

Cu-L: Yield 0.389 g (36.2%). Anal. Found: C, 56.65; H, 3.62; N, 9.44%. Calcd. for C₄₂H₃₂N₆O₈K₂Cu: C, 56.65; H, 3.62; N, 9.44%. IR (KBr, cm⁻¹) 532 (w), 596 (w), 691 (w), 769 (w), 784 (w), 1016 (w), 1114 (m), 1262 (w), 1334 (w), 1383 (m), 1407 (w), 1538 (m), 1610 (s, C=N), 1678 (w, C=O), 2359 (w), 2914 (w), 3435 (br, s, -OH).

Zn-L: Yield 0.389 g (21.9 %). Anal. Found: C, 59.04; H, 3.43; N, 8.91%. Calcd. for C₄₂H₃₀N₆O₆K₂Zn: C, 58.91; H, 3.30; N, 9.81%. IR (KBr, cm^-1) 513 (w), 596 (w), 686 (w), 786 (w), 1017 (w), 1112 (w), 1307 (w), 1385 (m), 1413 (w), 1548 (m), 1630 (s, C=N), 1721 (w, C=O), 2339 (w), 2966 (w), 3445 (br, s).

Elemental analyses (C, H and N) were performed using a Perkin-Elmer 2400 II CHNS/O analyzer at the Tokyo University of Science. Infrared (IR) spectra were recorded as KBr pellets on a JASCO FT-IR 4200 plus spectrophotometer in the range 4000–400 cm⁻¹ at 298 K. Electronic (UV-vis) spectra were obtained on a JASCO V-570 UV–vis–NIR spectrophotometer in the range 1500–200 nm at 298 K. Circular dichroism (CD) spectra were obtained on a JASCO J-820 spectropolarimeter in the range 900–250 nm at 298 K. Fluorescence spectra were recorded on a JASCO FP-6200 spectrophotometer at 298 K. Electrochemical (cyclic voltammetry, CV) measurements were carried out on a BAS SEC2000-UV/VIS and ALS2323 system with Ag/AgCl electrodes range of -0.50–0.80 V vs. Ag/Ag⁺. The conversion efficiency as DSSC cell was determined from J–V curves obtained under air mass 1.5 conditions at an illumination of 100 mW/cm² using an ADCMT 6241A DC voltage/current source/monitor according to the literature procedures.¹² The magnetic properties were investigated using a Quantum Design MPMS-XL superconducting quantum interference device (SQUID) magnetometer at an applied field of 1.0 T in the temperature range of 5-300 K. XPS measurement was carried out using Mg Kα source (10 kV, 25 mA) on a SHIMADZU ESCA3400.

Powder X-ray diffraction patterns of Fe-L and Cu-L were collected at 298 K with a Rigaku Smart Lab using Cu Kα source at the University of Tokyo and Pohang Light Source II 2D Supramolecular Crystallography Beamline (PLSII-2D-SMC), respectively. Powder of Cu-L was packed in the 0.5 mm diameter (wall thickness is 0.01 mm) capillary and the diffraction data measured transparency as Debye-Scherrer at 298 K respectively, with the 100 mm of detector distance in 10 sec exposures with synchrotron radiation (λ = 1.20007 Å) on an ADSC Quantum-210 detector at 2D SMC with a silicon (111) double crystal monochromator (DCM) at the Pohang Accelerator Laboratory, Korea. The PAL BL2D-SMDC program¹⁸ was used for data collection, and Fit2D program¹⁹ was used converted 2D to 1D pattern and wavelength and detector distance refinement. Rietveld analysis²⁰ was carried out with a Rigaku PDXL2 ver.2.2.1.0, commercially available program package by the following procedures: indexing, cell and space group determination, input composition and initial structural model from DFT, direct space method, addition of displacement parameters for non-hydrogen atoms, and refinement with hydrogen atoms under restraint (Fig. S1).

Crystallographic data (Table 1) of Fe-L and Cu-L have been deposited with the Cambridge Crystallographic Data Center as CCDC 1824200 and 1824206, respectively. These data can be obtained free of charge from the Cambridge Crystallographic Data Center via www.ccdc.cam.ac.uk/data_request/cif

Calculations of all complexes were performed using the Gaussian 09W software Revision D.02 (Gaussian, Inc.).²¹ The gas phase geometry optimizations were carried out using TD-DFT with B3LYP functional. The vertical excitation energy was calculated with the Lanl2dz for Fe, Cu and Zn with the 6-31+G(d) basis set for H, C, N and O method based on the singlet ground state geometry.

Fe-L crystalizes in Triclinic, space group P1 with Z = 1. As shown in Fig. 2 and Table 2, Fe-L affords a six-coordinated octahedral cis-[FeN₂O₄] coordination geometry with two axial water ligands, which is in agreement with the expected value of magnetic moment. For the chelate ligand, Fe1-O1, Fe1-O3, Fe1-N1, Fe1-N3 bond distances are ranging from 2.008 to 2.221 Å, while axial Fe1-O5 and Fe1-O8 bond distances are 2.221 and 2.103 Å, respectively. The chiral amine moiety adopts a λ configuration with torsion angle of C18-C10-C11-C12 = 64.01°.

Figure2.

Molecular structures of Fe-L.2H₂O showing selected atom labeling scheme. Hydrogen atoms and crystalline water are omitted for clarity.

Cu-L crystalizes in Triclinic, space group P1 with Z = 1. As shown in Fig. 3 and Table 3, Cu-L affords a six-coordinated elongated octahedral cis-[CuN₂O₄] coordination geometry with two axial water ligands. For the chelate ligand, Cu1-O1, Cu1-O3, Cu1-N1, Cu1-N3 bond distances are ranging from 1.985 to 2.151 Å, while axial Cu1-O3 and Cu1-O7 bond distances are 2.345 and 2.323 Å, respectively, which is expected due to Jahn-Teller distortion commonly. The chiral amine moiety adopts a λ configuration with torsion angle of C12-C11-C10-C18 = 61.07°. For both complexes, the other geometrical parameters are within a common range of the related compounds.¹²⁻¹⁵ Due to low crystal-linity, Zn-L could not be analyzed, though it may be also expected to be similar structures to Fe-L or Cu-L. Both complexes are aligned along the c axis without characteristic interactions (Fig. S2).

Figure3.

Molecular structures of Cu-L.2H₂O showing selected atom labeling scheme. Hydrogen atoms and crystalline water are omitted for clarity.

The UV-vis spectra (Fig. 4) appeared intense π-π* peaks at 368, 387, and 352 nm for Fe-L, Cu-L, and Zn-L, respectively, which are shifted to long wavelength region than the corresponding previous complexes without azomoiety (e.g. 322 nm for Fe(II) one).¹⁵ Due to π orbitals in the ligands, the bonds covered around 400 nm, in particular Fe-L exhibits a shoulder around 450-500 nm, for which DFT (based on an optimized structure, dipole moment of Fe-L is 1.74 D from Fe(II) ion to COO^- groups) exhibited HOMO-2→LUMO+1 and HOMO→ LUMO+4 (orbitals spreading in -COO^- groups) transitions at 376 and 449 nm, respectively. Therefore, owing to introduction of azo-moiety, three present complexes became to be useful dyes for DSSC in view of light absorption.

Figure4.

UV-vis spectra (in DMSO) for Fe-L, Cu-L, Zn-L, and previous Fe(II) complex (Fe-L’).¹⁵

Among the three complexes, only Zn-L showed fluorescence peak at 467 nm by excitation of 280 and 370 nm light (Fig. 5). The corresponding previous complex without azo-moiety exhibited at 465 nm,¹⁵ which is almost identical to Zn-L.

Figure5.

Fluorescence spectra (in DMSO) for Zn-L.

Table 4 lists HOMO-LUMO gap (E_g) and other electrochemical data obtained from CV measurement. For all complexes, LUMO exists in a higher level than the TiO₂ conduction band (-0.50 V), and HOMO exists in a lower level than the iodine (I^-/I₃^-) HOMO (+0.40 V). Therefore, it is conceivable that all complexes can inject electrons into TiO₂ and restore function as a dye for DSSC. Comparing with the previous Fe(II) complex,¹⁵ these differences are ascribed to central metal predominantly. Among them, the present Fe-L exhibited the best conversion efficiency (0.0055%). Comparing with N179 (0.0018%) under the same conditions (generally 7-10%²²⁻²³), some issues of assembling cells may significantly affect the low performance more than amount of dyes (see next section).

At first, sample was prepared on ITO substrates, The TiO₂ paste involving polyethylene glycol (molecular weight 2000) was coated on an indium doped tin-oxide (ITO) using spin-coat method. The ITO glass supported TiO₂ film (0.25 cm²) was then sintered at 723 K for 1 h. The electrode was immersed into solution of Zn-L (0.3 mM DMSO, 24 h) when the oven temperature was cooled to 313 K. Adsorption of Zn-L complex on TiO₂ surface was confirmed with XPS, appearance of not only 1024.6 eV (Zn2p_3/2) and 1047.6 eV (Zn2p_1/2) but also shift to 460.9 eV (Ti2p_3/2) 466.6 eV (Ti2p_1/2).

The amount of Fe-L adsorbed on TiO₂ was increased up to 61% after linearly polarized UV irradiation, which was found for the first time. Dye loading amounts were 2.33, 3.43, and 3.75×10^-7 mol/cm² for before and after natural or linearly polarized UV light (<350 nm) for 10 min, respectively. Since TiO₂ is a porous material (Fig. 6), it is considered that the amount of adsorption increased due to molecules entering the space of TiO₂ by irradiating polarized UV light to the dye and adsorbing it while trans to cis-photoisomerization and also reorientation towards anisotropic alignment (Weigert effect¹¹) and mutual aggregation.²⁴

Figure6.

Schematic representation of increasing adsorption amounts of complexes by polarized UV light irradiation. (blue circles) initial trans-azo-dye (red circles) aligned cis-azo-dye into porous surface.